Computer Architecture s Changing Definition

Transcription

1 Computer Architecture s Changing Definition 1950s Computer Architecture Computer Arithmetic 1960s Operating system support, especially memory management 1970s to mid 1980s Computer Architecture Instruction Set Design, especially ISA appropriate for compilers Vector processing and shared memory multiprocessors 1990s Computer Architecture Design of CPU, memory system, I/O system, Multi-processors, Networks Design for VLSI 2000s Computer Architecture: Special purpose architectures, Functionally reconfigurable, Special considerations for low power/mobile processing, highly parallel structures 1

2 Levels of Representation High Level Language Program Compiler Assembly Language Program Assembler Machine Language Program temp = v[k]; v[k] = v[k+1]; v[k+1] = temp; lw $15, 0($2) lw $16, 4($2) sw $16, 0($2) sw $15, 4($2) Machine Interpretation Control Signal Spec ALUOP[0:3] <= InstReg[9:11] & MASK 2

3 Levels of Abstraction Application Programming System Programming Computer Design Processor Graphical Interface Application Libraries Operating System Programming Language Assembler Language Instruction Set Architecture - Machine Language Fabrication Firmware Datapath and Control Logic Design Circuit Design Semiconductors Materials IO System Microprogramming Digital Design Circuits and devices 3

4 This Course Focuses on General Purpose Processors A general-purpose computer system Uses a programmable processor Can run any application Potentially optimized for some class of applications Common names: CPU, DSP, NPU, microcontroller, microprocessor Unified main memory For both programs & data Von Neumann computer Busses & controllers to connect processor, memory, IO devices Processor Input Control Memory Datapath Output MIT Whirlwind, 1951 Computers are pervasive servers, standalone PCs, network processors, embedded processors, 4

5 Metrics of Efficiency - Examples Desktop computing Examples: PCs, workstations Metrics: performance (latency), cost, time to market Server computing Examples: web servers, transaction servers, file servers Metrics: performance (throughput), reliability, scalability Embedded computing Examples: microwave, printer, cell phone, video console Metrics: performance (real-time), cost, power consumption, complexity 5

6 Moore s Law Moore s Law - The observation made in 1965 by Gordon Moore, co-founder of Intel, that the number of transistors per square inch on integrated circuits had doubled every year since the integrated circuit was invented Moore predicted that this trend would continue for the foreseeable future In subsequent years, the pace slowed down a bit, but data density has doubled approximately every 18 months, and this is the current definition of Moore's Law, which Moore himself has blessed Most experts, including Moore himself, expect Moore's Law to hold for at least another two decades 6

7 Bits DRAM Evolution size Year Capacity Access Kb 250 ns Kb 220 ns Mb 190 ns Mb 165 ns Mb 145 ns Mb 120 ns Mb 100 ns Gb 80 ns Year 7

8 Performance Perspectives Purchasing perspective Given a collection of machines, which has the - Best performance? - Least cost? - Best performance / cost? Design perspective Faced with design options, which has the - Best performance improvement? - Least cost? - Best performance / cost? Both require basis for comparison metric for evaluation 8

9 Definitions Performance is typically in units-per-second bigger is better If we are primarily concerned with response time performance = 1 execution_time " X is n times faster than Y" means ExecutionTime ExecutionTime y x Performance Performance x y n 9

10 Benchmarks Micro-benchmarks Measure one performance dimension - Cache bandwidth - Memory bandwidth - Procedure call overhead Insight into the underlying performance factors Not a good predictor of application performance Macro-benchmarks Application execution time - Measures overall performance, but on just one application - Need application suite 10

11 Why Do Benchmarks? How we evaluate differences Different systems Changes to a single system Provide a target Benchmarks should represent large class of important programs Improving benchmark performance should help many programs For better or worse, benchmarks shape a field Good ones accelerate progress good target for development Bad benchmarks hurt progress help real programs v sell machines/papers? Inventions that help real programs don t help benchmark 11

12 Metrics of Processor Performance Application Programming Language Compiler ISA Datapath Control Function Units Transistors Wires Pins Seconds per program Useful Operations per second (millions) of Instructions per second MIPS (millions) of (FP) operations per second MFLOP/s Megabytes per second Cycles per second (clock rate) Instructions perform read/write/load/store operations in a series of clock cycles 12

13 Organizational Trade-offs Cycles-Per-Instruction (CPI) Application Programming Language Compiler ISA Datapath Control Function Units Transistors Wires Pins is a useful design measure relating the Instruction Set Architecture with the Implementation of that architecture, and the program measured eg Can we handle more Instruction Mix CPI Cycle Time than one instruction in one clock cycle? 13

14 Processor Cycles Clock Combination Logic Cycle Most computers have fixed, repeating clock cycles 14

15 CPU Performance CPI Clock rate 15

16 Cycles Per Instruction (Throughput) Cycles per Instruction CPI = (CPU Time * Clock Rate) / Instruction Count = Cycles / Instruction Count CPU time Cycle Time Instruction Frequency n CPI j I j 1 j Ij is the count of instruction-j in the total instruction set CPI n CPI j 1 j F j where F j I Instruction j Count 16

17 Example Typical Mix Op Freq Cycles CPI ALU 50% 1 05 Load 20% 5 10 Store 10% 3 03 Branch 20% Remember: CPI is F x #cycles How much faster would the machine be if a better data cache reduced the average load time to 2 cycles? Load 20% x 2 cycles = 4 Total CPI Relative performance is 22 / 16 = 138 faster How does this compare with reducing the branch instruction to 1 cycle? Branch 20% x 1 cycle = 2 Total CPI Relative performance is 22 / 20 =

18 Evaluating Instruction Sets and Implementation Design-time metrics: Can it be implemented, in how long, at what cost? Can it be programmed? Ease of compilation? Static Metrics: How many bytes does the program occupy in memory? Dynamic Metrics: How many instructions are executed? How many bytes does the processor fetch to execute the program? How many clocks are required per instruction? Best Metric: Time to execute the program! CPI NOTE: Depends on instructions set, processor organization, and compilation techniques Inst Count Cycle Time 19

19 Clocking Methodology Clock Combination Logic All storage elements are clocked by the same clock edge Inputs are updated at each clock tick All outputs MUST be stable before the next clock tick 21

20 Critical Path & Cycle Time Clock Instructions follows various paths in these logic gates Critical path: the slowest path between any two storage devices Cycle time is a function of the critical path 22